Intro to Data
Visualization

Carlos Matos // ISPUP // November 2023

Data Visualization

is any graphical representation of information and data.

Data Visualization

helps to amplify cognition, gain insights, discover, explain, and make decisions.

Importance of visualizing data

Figures will often carry the weight of your arguments.

They need to be clear, attractive, and convincing.
A good visualization can be mean the difference between success and failure
- highly influential vs obscure paper
- grant or contract won or lost
- help managers make informed decisions
- Getting your point across!

Calculations do NOT replace graphs

13 datasets with similar x and y means, standard deviations and x-y correlation

library(datasauRus)
library(kableExtra)
datasaurus_dozen %>% 
    group_by(dataset) %>% 
    summarize(
      mean_x   = mean(x),
      mean_y   = mean(y),
      sd_x     = sd(x),
      sd_y     = sd(y),
      corr_x_y = cor(x, y)
    ) %>% 
  mutate(across(where(is.numeric), 
                scales::number,
                accuracy = 0.1))

# A tibble: 13 × 6
   dataset    mean_x mean_y sd_x  sd_y  corr_x_y
   <chr>      <chr>  <chr>  <chr> <chr> <chr>   
 1 away       54.3   47.8   16.8  26.9  -0.1    
 2 bullseye   54.3   47.8   16.8  26.9  -0.1    
 3 circle     54.3   47.8   16.8  26.9  -0.1    
 4 dino       54.3   47.8   16.8  26.9  -0.1    
 5 dots       54.3   47.8   16.8  26.9  -0.1    
 6 h_lines    54.3   47.8   16.8  26.9  -0.1    
 7 high_lines 54.3   47.8   16.8  26.9  -0.1    
 8 slant_down 54.3   47.8   16.8  26.9  -0.1    
 9 slant_up   54.3   47.8   16.8  26.9  -0.1    
10 star       54.3   47.8   16.8  26.9  -0.1    
11 v_lines    54.3   47.8   16.8  26.9  -0.1    
12 wide_lines 54.3   47.8   16.8  26.9  -0.1    
13 x_shape    54.3   47.8   16.8  26.9  -0.1

Calculations do NOT replace graphs

Fundamentals of data viz

What do a heatmap, a pie chart and a line chart have in common?

Fundamentals of data viz

→ All data visualizations map data values into quantifiable features of the resulting graphic

→ We refer to these features as aesthetics

“Carte figurative des pertes successives en hommes de l’Armée qu’Annibal conduisit d’Espagne en Italie en traversant les Gaules (selon Polybe)” (top) and “Carte figurative des pertes successives en hommes de l’Armée Française dans la campagne de Russie 1812–1813” (bottom) by Charles Joseph Minard (1869)

Mapped Dimensions

Nº of Napoleon’s troops
distance
temperature
latitude
longitude
direction of travel
location relative to specific dates

Common aesthetics

Common aesthetics in data visualization

All plots have

Data + Mapping

ggplot2

Intro to {ggplot2}

Has an underlying grammar
Easy to combine multiple datasets in the same plot
Solutions become more intuitive as we get to know the grammar
Made to work iteratively: start with a raw data layer and add annotations and statistical summaries as you go
Default graphics are quite good (publication-ready)

Anatomy of a ggplot

Five mapping components:

Layer: What we actually see in the plot. Geometric elements (“geoms”. e.g. points, lines, …) or statistical transformations (“stats”. e.g. counting, binning, fitting a model)
Scales: Convert the values in the data space to values in the aesthetic space
Coordinate system: Usually cartesian, but other are available (e.g. polar)
Facet: Subset displays
Theme: Control finer points of the visualization

Anatomy of a ggplot

Three of which are key components that exist in every plot:

→ Data

→ Mapping

→ Geom

Anatomy of a ggplot

The overall template of a ggplot looks like this

ggplot(data = DATA) +
    GEOM_FUNCTION(mapping = aes(MAPPINGS))

Where:
- DATA is your dataset
- Example GEOM_FUNCTION are geom_point(), geom_col(), geom_histogram(), …
- Example MAPPINGS are x = year, y = life_expectancy, fill = death_rate, …

First worked example

#Preparing the data
library(gapminder)

gm_latest <- gapminder %>% 
  filter(year == max(year)) %>% 
  rename(GDPpc = gdpPercap,
         LE = lifeExp)

gm_latest

# A tibble: 142 × 6
   country     continent  year    LE       pop  GDPpc
   <fct>       <fct>     <int> <dbl>     <int>  <dbl>
 1 Afghanistan Asia       2007  43.8  31889923   975.
 2 Albania     Europe     2007  76.4   3600523  5937.
 3 Algeria     Africa     2007  72.3  33333216  6223.
 4 Angola      Africa     2007  42.7  12420476  4797.
 5 Argentina   Americas   2007  75.3  40301927 12779.
 6 Australia   Oceania    2007  81.2  20434176 34435.
 7 Austria     Europe     2007  79.8   8199783 36126.
 8 Bahrain     Asia       2007  75.6    708573 29796.
 9 Bangladesh  Asia       2007  64.1 150448339  1391.
10 Belgium     Europe     2007  79.4  10392226 33693.
# ℹ 132 more rows

First worked example

#Plot
ggplot(data = gm_latest) #Data

First worked example

#Plot
ggplot(data = gm_latest) + #Data
  aes(x = GDPpc) #Mapping

First worked example

#Plot
ggplot(data = gm_latest) + #Data
  aes(x = GDPpc, y = LE) #Mapping

First worked example

#Plot
ggplot(data = gm_latest) + #Data
  aes(x = GDPpc, y = LE) + #Mapping
  geom_point() #Layer

First worked example

#Plot
ggplot(data = gm_latest) + #Data
  aes(x = GDPpc, y = LE) + #Mapping (global)
  geom_point(aes(size = pop)) #Layer with mapping

First worked example

#Plot
ggplot(data = gm_latest) + #Data
  aes(x = GDPpc, y = LE) + #Mapping (global, for all geom)
  geom_point(aes(size = pop, color = pop)) #Layer with mapping (only for this geom)

Layer order

#Plot
ggplot(gm_latest, aes(GDPpc, LE)) + 
  geom_point(aes(size = pop, color = pop)) + 
  geom_hline(yintercept = 73, 
             color = "red", 
             alpha = 1) #After geom_point

#Plot
ggplot(gm_latest, aes(GDPpc, LE)) + 
  geom_hline(yintercept = 73, 
             color = "red", 
             alpha = 1) + #Before geom_point
  geom_point(aes(size = pop, color = pop))

{ggplot2} Theme elements

Taking the mtcars dataset as an example

mtcars %>% 
  ggplot(aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point(size = 4) +
  theme(text = element_text(size = 18))

{ggplot2} Theme elements

Theme elements of a ggplot

Interactive plots with plotly

#Store the plot in an object
my_plot <- mtcars %>% 
  ggplot(aes(x = mpg, y = hp, 
             color = factor(cyl))) +
  geom_point() 

#Call ggplotly with that object 
#as argument
plotly::ggplotly(my_plot)

Static maps with ggplot

#Get the data
world <- map_data("world") 

#Plot
world %>% 
  ggplot() +
  geom_polygon(aes(x = long, y = lat, 
                   group = group))  +
  theme(legend.position = "none") +
  theme_void() +
  coord_equal()

Dynamic maps with leaflet

For more info about leaflet check the website

library(leaflet)

leaflet(width = 1100, height = 300) %>% 
  addTiles() %>% 
  addPopups(lng= -8.616611357347235, 
            lat=41.14440087224056, 
            popup="ISPUP!")

Intro to Data Visualization

Data Visualizationis any graphical representation of information and data.

Data Visualizationhelps to amplify cognition, gain insights, discover, explain, and make decisions.

Importance of visualizing data

Calculations do NOT replace graphs

Calculations do NOT replace graphs

Calculations do NOT replace graphs

Fundamentals of data viz

Fundamentals of data viz

Common aesthetics

All plots have Data + Mapping

ggplot2

Intro to {ggplot2}

Anatomy of a ggplot

Anatomy of a ggplot

Anatomy of a ggplot

First worked example

First worked example

First worked example

First worked example

First worked example

First worked example

First worked example

Layer order

{ggplot2} Theme elements

{ggplot2} Theme elements

Interactive plots with plotly

Static maps with ggplot

Dynamic maps with leaflet

Exercises

Intro to Data
Visualization

Data Visualization

is any graphical representation of information and data.

Data Visualization

helps to amplify cognition, gain insights, discover, explain, and make decisions.

All plots have

Data + Mapping